Acute respiratory distress syndrome (ARDS) is a form of acute respiratory failure resulting from a variety of insults including sepsis, smoke inhalation and severe trauma. ARDS has a high mortality rate of 30-40%, which results in approximately 75,000 deaths per year. This exceeds the mortality due to breast or prostate cancer. Treatment of ARDS is based on supportive mechanical ventilation that is applied while the underlying cause of respiratory failure hopefully resolves. However, selecting appropriate ventilation parameters is difficult because of the conflicting requirements imposed by the inhomogeneous nature of lung injury in ARDS. Inspiratory pressures must be sufficiently low to avoid over-distention of the delicate parenchyma (volutrauma) while at the same time expiratory pressures must be high enough to prevent damage caused by the repetitive collapse (derecruitment) and reopening (recruitment) of airways and alveoli (atelectrauma). Volutrauma and atelectrauma can both lead to ventilator-induced lung injury (VILI) which is manifest as local accumulation of edema in the airspaces. This, in turn, leads to surfactant inactivation, increased tissue stress, and further VILI in a positive feedback mechanism that often leads to death. However, exactly how VILI begins within the lung tissue, and then develops over time, remains poorly understood. We hypothesize that edema and atelectasis begin locally in regions of high tissue stress and then propagate outward to consume the rest of the lung as a result of fluid-structure interactions. This is exacerbated during mechanical ventilation because ventilation heterogeneity amplifies the damage generated in local stress foci. We will test this hypothesis by using design- based stereology to quantify how the spatial distributions of edema and atelectasis change with time during the progression of VILI in mouse models of ARDS. These measurements will then inform the development of a computational model of an alveolar network that couples solid and fluid mechanics to determine how inhomogeneous edema alters microscale tissue stress and recruitment/derecruitment. The numerical model will be used to investigate potentially protective modes of mechanical ventilation, such as variable tidal volume ventilation, that avoid persistently concentrating stress in fixed regions of the lung tissue, as tends to occur with conventional regular ventilation. These studies will facilitate the development of novel protective ventilation strategies for ARDS and thereby help reduce mortality. The PI of this proposal has extensive experience with numerical modeling, animal experimentation, and organ-scale physiology. Complementary training in morphometric analysis will provide the PI with the skills necessary to quantify the micro-scale effects of lung injury, and to link these structural changes to lung function and injury progression using computational models. This program of study and research, together with the world-class research environment provided by the University of Vermont College of Medicine, will enable the PI to develop a career as an independent investigator applying bioengineering and computational methods to the study of lung disease.